Correct usage of Calibration from Image? (Vision Assistant)

[mathieu2011]

Hey there!

I am currently building a front end for an inspection device based on a line-camera moving on a 1D Axis.

What i'm having problems with is understanding the principles behind the image calibration modules. Basically i need the mm/px values in each direction to calculate the size/area of the material defects i'm trying to detect with the IMAQ search objects 2.vi (I am not using the Particle Analysis preset from vision since it seems to lack the option to limit the search to specific sizes)

--

What i originally intended to do was:
Acquire image of a calibration grid -> Image Calibration -> save calibration image as .png -> read out the mm to px conversion and statically multiply it with the pixel area output from IMAQ search objects 2.vi

However, since i got a nonlinear distortion i figured i probably can not just use the same mm/px conversion over the whole image. So what i did in addition to the above steps was:

Image correction of the calibration -> another Image Calibration to get the proper mm/px which should then be constant over the whole image assuming the image correction was accurate.

The problem with this method is, that my images are of a size around 8000x7000px, hence the image correction takes ages to apply within LabView (around a minute or more)

Am i overdoing it with the second calibration step after the image correction? Could i just leave the image correction out and apply the mm/px values from my first calibration to my acquired image and still get remotely accurate results?


Thanks in advance

Mathias

[Pixar]

Hi Mathias

Problem is performing a calibration with a large image size need a lot of performance. So when you can't decrease the resolution it will need that much time.

For calibration examples please have a look at this:

Nonlinear Calibration Example

Perspective Calibration Example

After a calibration all following steps should return their results in mm. The best is to create a script in the Vision Assistant and create a VI with it.

I think you can also learn the calibration once and after that use the IMAQ Learn Perspective Calibration VI for example.

Kind regards

[mathieu2011]

Thanks for the input.. seems like i will have to settle with the long calibration time then.